CONCLUSION
Data clustering has always been of importance in the field of computer science. With the technology shift from small,
scalable and manageable data towards big data has opened up many complex problems with even more complex yet efficient solutions. Big data video analytics is a growing research area. It has become more challenging with the increasing lengths and the variety of videos being uploaded every single second. In this paper we have analyzed the problem of extracting content information and generating classes based on that. K-means clustering and skeleton algorithm have been studied and combined to present a more appropriate and time efficient algorithm for the growing big data video analytics and storage. In future we intend on working with videos retrieval from the cloud sets and more precisely for extracting video information for both business and computational logic. Video retrieval and analysis for business intelligence and defense sector has recently taken an electric spike. These areas focus
on observing human behavior, the behavior of actions and associated reactions to come towards a single conclusion of people oriented marketing and sales. For defense sector multiple surveillance and other videos can be analyzed to
generate a more specific algorithm for war planning and security orientations